Scheuermann Disease

Abstract

Observational population-based study. To determine the prevalence of radiographical Scheuermann disease in a Dutch population and evaluate the consistency of diagnostic criteria. Scheuermann disease is a form of osteochondrosis characterized by increased posterior rounding of the thoracic spine with structural vertebral deformity. Different expert opinion-based radiological criteria exist, yet these have not been validated. The prevalence in the general population reported ranged from 1% to 10%. Lateral spine radiographs of 2753 Rotterdam Study participants (aged 45-89 yr) were assessed for Scheuermann disease using Sørensen and Sachs' radiographical criteria in 2 phases. Cohen κ statistics were calculated for interrater agreement. Prevalence estimates were calculated and sex differences were tested with Pearson χ test. We evaluated whether varying the kyphosis angle criterion would change the prevalence estimate. A total of 677 (24.6%) individuals had endplate irregularities and 140 (5.1%) individuals had vertebral wedging. Abnormalities were significantly more prevalent among males (P < 0.05). The interrater agreement κ statistics were 78.8% for vertebral wedging and 79.4% for endplate irregularity. A total of 127 individuals had both criteria, of which 111 had a kyphosis angle greater than 45°, resulting in a prevalence of 4.0% (95% confidence interval [CI]: 3.3%-4.7%). The disease prevalence was 4.5% in males versus 3.6% in females, yet this difference was not statistically significant (P = 0.23). Adjustment of the kyphosis angle criterion from 45° to 40° or 35° increased the number of cases marginally, corresponding to prevalence estimates not significantly different from the estimates using original criteria (4.2% [95% CI: 3.3%-4.7%] and 4.4% [95% CI: 3.6%-5.2%]). Our results revealed a prevalence of 4.0% of radiographical Scheuermann disease in Dutch individuals aged 45 years and older. Although there is no current "gold standard" for the radiographical definition, standardized scoring of independent features resulted in substantial interobserver agreement, and different applications of diagnostic criteria did not significantly alter the classification.